22NO

Technical details

Show code

library(GeoPressureR)
library(leaflet)
library(leaflet.extras)
library(raster)
library(dplyr)
library(ggplot2)
library(kableExtra)
library(plotly)
library(GeoLocTools)
setupGeolocation()
knitr::opts_chunk$set(echo = FALSE)
load(paste0("../data/1_pressure/", params$gdl_id, "_pressure_prob.Rdata"))
load(paste0("../data/2_light/", params$gdl_id, "_light_prob.Rdata"))
load(paste0("../data/3_static/", params$gdl_id, "_static_prob.Rdata"))
load(paste0("../data/4_basic_graph/", params$gdl_id, "_basic_graph.Rdata"))

Settings used

All the results produced here are generated with (1) the raw geolocator data, (2) the labeled files of pressure and light and (3) the parameters listed below.

Show code

kable(gpr) %>% scroll_box(width = "100%")

gdl_id	crop_start	crop_end	thr_dur	extent_N	extent_W	extent_S	extent_E	map_scale	map_max_sample	map_margin	prob_map_s	prob_map_thr	shift_k	kernel_adjust	calib_lon	calib_lat	calib_1_start	calib_1_end	calib_2_start	calib_2_end	calib_2_lon	calib_2_lat	prob_light_w	thr_prob_percentile	thr_gs	RingNo	scientific_name	common_name	mass	wing_span	Color
22NO	2019-01-24	2019-09-29	6	17	9	-25	38	5	300	30	1.2	0.9	0	1.4	28.77491	-22.72541	2019-01-24	2019-04-17	NA	NA	NA	NA	0.1	0.95	120	NA	Halcyon senegaloides	Woodland Kingfisher	NA	NA	#E9FF70

Pressure timeserie

The labeling of pressure data is illustrated with this figure. The black dots indicates the pressure datapoint not considered in the matching. Each stationary period is illustrated by a different colored line.

Show code

pressure_na <- pam$pressure %>%
  mutate(obs = ifelse(isoutliar | sta_id == 0, NA, obs))
p <- ggplot() +
  geom_line(data = pam$pressure, aes(x = date, y = obs), colour = "grey") +
  geom_point(data = subset(pam$pressure, isoutliar), aes(x = date, y = obs), colour = "black") +
  # geom_line(data = pressure_na, aes(x = date, y = obs, color = factor(sta_id)), size = 0.5) +
  geom_line(data = do.call("rbind", shortest_path_timeserie) %>% filter(sta_id > 0), aes(x = date, y = pressure0, col = factor(sta_id))) +
  theme_bw() +
  scale_colour_manual(values = pam$sta$col) +
  scale_y_continuous(name = "Pressure(hPa)")

ggplotly(p, dynamicTicks = T) %>% layout(showlegend = F)

Pressure calibration

Show code

sp_pressure = do.call("rbind", shortest_path_timeserie) %>% filter(sta_id > 0)

sta_plot <- which(difftime(pam$sta$end,pam$sta$start,unit="days")>3)

par(mfrow=c(2,3))
for (i in seq_len(length(sta_plot))){
  i_s = sta_plot[i]
  pressure_s = pam$pressure %>% 
    filter(sta_id==i_s & !isoutliar)
  
    err <- pressure_s %>% left_join(sp_pressure, by="date") %>% 
      mutate(
        err = obs-pressure-mean(obs-pressure)
      ) %>% .$err
    
    hist(err, freq = F, main = paste0("sta_id=",i_s, " | ",nrow(pressure_s)," dtpts | std=",round(sd(err),2)))
   xfit <- seq(min(err), max(err), length = 40) 
    yfit <- dnorm(xfit, mean = mean(err), sd = sd(err)) 
    lines(xfit, yfit, col = "red")
}

Light

Show code

raw_geolight <- pam$light %>%
  transmute(
    Date = date,
    Light = obs
  )
lightImage(tagdata = raw_geolight, offset = 0)
tsimagePoints(twl$twilight,
  offset = 0, pch = 16, cex = 1.2,
  col = ifelse(twl$deleted, "grey20", ifelse(twl$rise, "firebrick", "cornflowerblue"))
)
abline(v = gpr$calib_2_start, lty = 1, col = "firebrick", lwd = 1.5)
abline(v = gpr$calib_1_start, lty = 1, col = "firebrick", lwd = 1.5)
abline(v = gpr$calib_2_end, lty = 2, col = "firebrick", lwd = 1.5)
abline(v = gpr$calib_1_end, lty = 2, col = "firebrick", lwd = 1.5)

Show code

hist(z, freq = F)
lines(fit_z, col = "red")

The probability map resulting from light data alone can be seen below.

Show code

li_s <- list()
l <- leaflet(width = "100%") %>%
  addProviderTiles(providers$Stamen.TerrainBackground) %>%
  addFullscreenControl()
for (i_r in seq_len(length(light_prob))) {
  i_s <- metadata(light_prob[[i_r]])$sta_id
  info <- pam$sta[pam$sta$sta_id == i_s, ]
  info_str <- paste0(i_s, " | ", info$start, "->", info$end)
  li_s <- append(li_s, info_str)
  l <- l %>% addRasterImage(light_prob[[i_r]], opacity = 0.8, colors = "OrRd", group = info_str)
}
l %>%
  addCircles(lng = gpr$calib_lon, lat = gpr$calib_lat, color = "black", opacity = 1) %>%
  addLayersControl(
    overlayGroups = li_s,
    options = layersControlOptions(collapsed = FALSE)
  ) %>%
  hideGroup(tail(li_s, length(li_s) - 1))

Light vs Pressure

We can compare light and pressure location at long stationary stopover (>5 days). By assuming the best match of the pressure to be the truth, we can plot the histogram of the zenith angle and compare to the fit of kernel density at the calibration site.

Show code

 raw_geolight <- pam$light %>%
    transmute(
      Date = date,
      Light = obs
    )
 dur <- unlist(lapply(pressure_prob, function(x) difftime(metadata(x)$temporal_extent[2],metadata(x)$temporal_extent[1], units = "days" )))
  long_id <- which(dur>5)

par(mfrow = c(2, 3))
for (i_s in long_id){
  twl_fl <- twl %>%
    filter(!deleted) %>%
    filter(twilight>shortest_path_timeserie[[i_s]]$date[1] & twilight<tail(shortest_path_timeserie[[i_s]]$date,1))
  sun <-  solar(twl_fl$twilight)
  z_i <- refracted(zenith(sun, shortest_path_timeserie[[i_s]]$lon[1], shortest_path_timeserie[[i_s]]$lat[1]))
  hist(z_i, freq = F, main = paste0("sta_id=",i_s, " | ",nrow(twl_fl),"twls"))
  lines(fit_z, col = "red")
  xlab("Zenith angle")
}

Similarly, we can plot the line of sunrise/sunset at the best match of pressure (yellow line) and compare to the raw and labeled light data.

Show code

  lightImage(
    tagdata = raw_geolight,
    offset = gpr$shift_k / 60 / 60
  )
  tsimagePoints(twl$twilight,
                offset = gpr$shift_k / 60 / 60, pch = 16, cex = 1.2,
                col = ifelse(twl$deleted, "grey20", ifelse(twl$rise, "firebrick", "cornflowerblue"))
  )
  for (ts in shortest_path_timeserie){
    if (!is.null(ts)){
      twl_fl <- twl %>%
      filter(twilight>ts$date[1] & twilight<tail(ts$date,1))
      if (nrow(twl_fl)>0){
      tsimageDeploymentLines(twl_fl$twilight,
                             lon = ts$lon[1], ts$lat[1],
                             offset = gpr$shift_k / 60 / 60, lwd = 3,col = adjustcolor("orange", alpha.f = 0.5))
        
      }
    }
  }

GeoPressureViz

To visualize the path on GeoPressureViz, you will need to also load the pressure and light probability map and align them first with the code below.

Show code

sta_marginal <- unlist(lapply(static_prob_marginal, function(x) raster::metadata(x)$sta_id))
sta_pres <- unlist(lapply(pressure_prob, function(x) raster::metadata(x)$sta_id))
sta_light <- unlist(lapply(light_prob, function(x) raster::metadata(x)$sta_id))
pressure_prob <- pressure_prob[sta_pres %in% sta_marginal]
light_prob <- light_prob[sta_light %in% sta_marginal]

The code below will open with the shortest path computed with the graph approach.

Show code

geopressureviz <- list(
  pam_data = pam,
  static_prob = static_prob,
  static_prob_marginal = static_prob_marginal,
  pressure_prob = pressure_prob,
  light_prob = light_prob,
  pressure_timeserie = shortest_path_timeserie
)
save(geopressureviz, file = "~/geopressureviz.RData")

shiny::runApp(system.file("geopressureviz", package = "GeoPressureR"),
  launch.browser = getOption("browser")
)

Stationay period information

Show code

pam$sta %>% mutate(duration = difftime(end,start,units="days")) %>% kable()

start	end	sta_id	col	duration
2019-01-24 00:00:00	2019-04-16 16:50:00	1	#1B9E77	82.7013889 days
2019-04-17 04:10:00	2019-04-17 17:00:00	2	#D95F02	0.5347222 days
2019-04-17 23:15:00	2019-04-18 17:25:00	3	#7570B3	0.7569444 days
2019-04-18 23:05:00	2019-04-19 19:50:00	4	#E7298A	0.8645833 days
2019-04-19 21:00:00	2019-04-20 21:55:00	5	#66A61E	1.0381944 days
2019-04-20 22:15:00	2019-05-02 20:25:00	6	#E6AB02	11.9236111 days
2019-05-03 01:10:00	2019-05-03 19:40:00	7	#A6761D	0.7708333 days
2019-05-03 21:50:00	2019-05-06 20:10:00	8	#666666	2.9305556 days
2019-05-06 22:50:00	2019-05-07 21:50:00	9	#1B9E77	0.9583333 days
2019-05-07 22:55:00	2019-05-14 00:35:00	10	#D95F02	6.0694444 days
2019-05-14 01:40:00	2019-05-15 19:10:00	11	#7570B3	1.7291667 days
2019-05-15 21:45:00	2019-05-16 22:15:00	12	#E7298A	1.0208333 days
2019-05-16 22:40:00	2019-05-18 01:15:00	13	#66A61E	1.1076389 days
2019-05-18 01:25:00	2019-05-18 02:05:00	14	#E6AB02	0.0277778 days
2019-05-18 02:10:00	2019-05-24 23:20:00	15	#A6761D	6.8819444 days
2019-05-25 01:40:00	2019-05-25 23:10:00	16	#666666	0.8958333 days
2019-05-25 23:20:00	2019-05-27 20:35:00	17	#1B9E77	1.8854167 days
2019-05-27 22:35:00	2019-05-31 19:20:00	18	#D95F02	3.8645833 days
2019-05-31 21:15:00	2019-06-04 18:05:00	19	#7570B3	3.8680556 days
2019-06-05 03:45:00	2019-06-05 23:20:00	20	#E7298A	0.8159722 days
2019-06-06 01:15:00	2019-06-09 19:45:00	21	#66A61E	3.7708333 days
2019-06-09 21:25:00	2019-06-10 23:15:00	22	#E6AB02	1.0763889 days
2019-06-11 00:55:00	2019-06-18 23:15:00	23	#A6761D	7.9305556 days
2019-06-19 01:45:00	2019-06-19 19:10:00	24	#666666	0.7256944 days
2019-06-20 00:05:00	2019-06-20 22:35:00	25	#1B9E77	0.9375000 days
2019-06-21 00:35:00	2019-06-21 21:10:00	26	#D95F02	0.8576389 days
2019-06-21 22:10:00	2019-06-22 00:50:00	27	#7570B3	0.1111111 days
2019-06-22 01:10:00	2019-06-22 01:30:00	28	#E7298A	0.0138889 days
2019-06-22 01:35:00	2019-06-22 17:55:00	29	#66A61E	0.6805556 days
2019-06-22 22:15:00	2019-06-23 20:15:00	30	#E6AB02	0.9166667 days
2019-06-23 21:45:00	2019-06-23 23:45:00	31	#A6761D	0.0833333 days
2019-06-24 00:35:00	2019-06-25 19:35:00	32	#666666	1.7916667 days
2019-06-25 20:25:00	2019-06-26 21:45:00	33	#1B9E77	1.0555556 days
2019-06-26 22:10:00	2019-06-27 20:15:00	34	#D95F02	0.9201389 days
2019-06-27 22:10:00	2019-07-01 01:40:00	35	#7570B3	3.1458333 days
2019-07-01 01:55:00	2019-07-01 20:50:00	36	#E7298A	0.7881944 days
2019-07-01 21:10:00	2019-07-03 02:05:00	37	#66A61E	1.2048611 days
2019-07-03 02:35:00	2019-07-04 23:55:00	38	#E6AB02	1.8888889 days
2019-07-05 00:40:00	2019-07-07 00:15:00	39	#A6761D	1.9826389 days
2019-07-07 01:05:00	2019-07-08 01:40:00	40	#666666	1.0243056 days
2019-07-08 02:25:00	2019-09-28 23:55:00	41	#1B9E77	82.8958333 days